series capacitor compensation requirements

8/12/2019 Series Capacitor Compensation Requirements

1/35

4 Series Capacitor Compensation Requirements

4 1

Purposes and Benefits of Series Capacitor Compensation

As noted previously, transmission lines inherently have an inductive reactance that is in series

with flow of current between the source and the load. This impedance is responsible for a

significant portion of the voltage drops in the transmission systems and is proportional to the

length of the transmission lines. Transmission line designers will attempt to keep the line

reactance as low as possible because it provides significant benefit to keeping the system

tightly connected - that is, it keeps the sources of generation electrically closer to the load.

Higher transmission voltages can effectively reduce the influence of line reactance, not only

because of differences in the line designs, but primarily because significantly reduced levels of

current flow for a given amount of power being transmitted resulting in correspondingly reduced

levels of voltage drop along the line. Based on a cost-effectiveness analysis, 345kV was

selected as the appropriate voltage level for the CTP.

The long distances associated with the transmission of the wind energy from the CREZ to the

load centers results in several long transmission lines between the various system buses. Some

of these are so long, that system stability is impacted and it becomes necessary to find a way to

reduce the reactance associated with these lines. One method is to increase the number of

circuits along the critical paths, but this is not economically desirable - particularly considering

that the circuits will be under utilized for the amount of power to be transferred. A well known

and understood method is to compensate a portion of the series line inductive reactance with a

series capacitor. At normal system operating frequencies and from the perspective of total line

reactance, this is the same as reducing the line length in proportion to the level of series

compensation.

Ideally,

the total line reactance would be zero

( or at least, very

low) suggesting

that 100%

compensation is desirable. However, there are other design considerations, such as the

voltages along the length of the l ine and reson ances that can res ult in severe interactions w ith

conventional thermal generat ion,

which

mu st also be considered in select ing the opt im al level of

ser ies com pensat ion. Several of these design issu es are discusse d in m ore deta i l in Sect ions

4.3 and 4.4.

In order to address som e of these design issu es, series capacitors can be des igned

with m ult iple sm aller se gmen ts placed at dif ferent locations along the

line. Typically only one or

two segments are used.

In selecting the final series compensation levels for the CREZ transmission lines, multiple

issues including line voltage profiles, voltage stability, system angular stability and

subsynchronous resonances were evaluated. Several compensation levels up to

75% were

considered, but, driven

primarily by the TSP's

proposed l ine des ign cri teria

ERCOT

supported

the f indings that com pens ation levels of approximate ly

50% represented a good comprom ise

ABB Grid Systems Consulting

'28

ERCOT

000050


2/35

among competing constraints. The actual percentage of series compensation will vary slightly

depending on the final length of the associated line and TSP implementation as a result of

procurement.

4.2

Study Locations of Series apacitors

Based on interim s tudy results, several changes were m ade to the CTP

during the course of the

study. The

f inal series com pensated l ines and se ries capacitor locations shown in Table 4.2-1

were us ed for the study.

Table 4.2-1: CREZ Series Capacitor Locations as Studied

TS P

Line

Circuit

#

Segmen t

#

Study Series Capaci tor Locat ion

CTT

Silverton-Tesla

Mid-line

2 Mid-line

ETT

Edith Clarke- Clear Crossing North

Mid-line

2 Mid-line

Dermott - Clear Crossing West

Mid-line

2

Mid-line

Big Hill - Kendall

Mid-line at Edison

2

Midway between Big Hill and Edison

2

Mid-line at Edison

2

2 Midway betw een Big Hill and Edison

ONCOR

Wil low Creek- C lear Crossing East Clear Crossing East

2 Clear Crossing East

Lone

W. Shackelford - Sam Sw itch

Rom ney 1(-1/3 from W. Shackelford)

Star

2 Kopperl 1 (-1 /3 from Sam Switch)

W. S hackelford - Navarro

2

Romney 2

2 2

Kopper l2

The locations of the series capacitor segments along the length of these lines as studied were

provided by ERCOT and the TSPs. The ultimate locations on the lines will be established by the

TSPs based on maintenance needs, line design criteria and similar considerations. The

locations on the lines will not influence the reactive compensation requirements.

The comprehensive reactive compensation plan developed for the CREZ initial build was

developed assuming the series compensation levels indicated on the lines listed. It is therefore

assumed that the series compensation is installed as an integral part of the initial build of the

CREZ transmission system.

4.3

Study Approach to Determining Series Capacitor Requirements

4.3. 9

Series Capacitor Technology

Series compensation to reduce the effective impedance of a transmission line can be

accomplished by putting a capacitor bank in series with the line. This series capacitor bank will

E3800-PR-00

CREZ Reactive Power Compensation Study

29

0 005i


3/35

be installed on a platform that is insulated against the full line voltage since all of the equipment

will be operating at the line voltage potential.

A typical series capacitor bank consists of the arrangement shown in Figure 4.3-1. The actual

design of these series capacitor banks is subject to detail design studies considering the actual

network data and system requirements. The main components of the series compensation

include the series capacitor bank, the MOV overvoltage protection, a bypass gap and/or bypass

switch. All of the components, except the disconnectors and the bypass switch, are normally on

the capacitor platform.

Bypass

Isolating

disconnector

isconnector\

\

Isolating

\ disconnector

I I--

Discharge current

l imit ing reactor

M O V

Bypass gap

Bypass switch (breaker)

IF-

Platform structure

Figure 4.3-1: Series capacitor bank main components

During fault conditions, series capacitor units are generally subjected to overvoltages which are

related to the fault current levels. When, like in the CREZ system, the series capacitors are at

substations with limited transformations and long transmission lines, the highest fault currents -

and therefore the highest overvoltages - are expected with three-phase faults. When a station

has large transformers and shorter lines, it is possible for single-phase faults to result in higher

fault currents. The fault current levels and the resulting overvoltages on the series capacitors

need to be confirmed during the design stage.

I

ABB Grid Systems Consult ing

30

ERCOT

0.00052


4/35

Fault related overvoltages may persist until the fault is cleared by opening of the line circuit

breakers to the faulted circuit element. Modern series capacitor banks use highly non-linear

Metal Oxide Varistors (MOV) to limit the voltage across the series capacitor to a desired

protective level,

which typically ranges between 2.0 and 2.5 times the voltage across the

capacitor at the rated bank current. When limiting the voltage across the series capacitor to the

protective level during fault conditions, the MOV must conduct the excess fault current and

thereby absorb energy. The MOV energy is kept within the MOV's absorption capability by

bypassing the parallel capacitor/MOV combination using two devices. The first is a very fast

acting device called a triggered spark gap. After the spark gap is triggered, a slower acting

bypass breaker will close. From a system performance point of view, overvoltage protection

bypasses the series capacitor, thereby increasing the impedance of the circuit. This may, in

turn, adversely impact network stability. The effect is not significant for faults that occur on the

line section in which the series capacitors are located (i.e. internal faults), because the line

section containing the series capacitor bank is eventually removed from service to allow fault

clearing.

For faults not on the same line as the series capacitor (i.e. external faults) the impact on system

stability can be significant. Therefore, whichever type of overvoltage protection scheme is

adopted, it is usually designed so that the capacitor bank is not bypassed during external faults.

Protective bypassing is restricted by design to act only for the more severe internal faults

exceeding the specified energy and fault current.

Series capacitor compensation includes a microprocessor based control and protection system

and the inputs are the currents measured at several points on the capacitor platform.

The main system requirements for rating the series capacitor banks are:

Rated capacitor reactive impedance (ohm s)

Continuous capacitor current requirem ents

(

amperes)

30 minute overload current requirem ents

(

amperes)

.

Maximum swing current fol lowing system disturbances

Maximu m fault current for external faults

Maximu m fault current for internal faults

The rated reactive power and rated bank

(

series voltage) are determined based on the f irst two

items. The MOV ratings are determ ined based on the fault currents.

As m entioned previously, the total series capacitor impedance for each compe nsated CR EZ l ine

was se lected to be approxim ately

50% of

the l ine im pedance base d on the analysis of m ult iple

issues. For those l ines wi th two segm ents, each segm ent was approximately

25% of the line

impedance.

E3800-PR-00


31

53


5/35

4.3.2 Line Voltage Profiles

The decision to limit the total compensation of the series compensated CREZ lines to 50% was

based primarily on the profile of the voltages along the length of the lines. The TSPs' design

criteria limit the voltage at any point on the line to 105% under normal conditions and 110% on

contingencies for up to 30 minutes. In order to meet the voltage criteria, the series

compensation had to be limited to 50% and placed in the middle of the lines except for the Clear

Crossing-Willow Creek lines for which a similar action is recommended.

Although some initial study work considered higher compensation levels, which showed

improved system performance, these higher levels of compensation were not able to meet the

voltage limit criteria. However, higher levels of series compensation and/or locations at the end

of lines could be accommodated if line designs allow for higher line voltages. Figure 4.3-2

below is an example of the line voltage profile for series capacitors at the end of the line and in

the middle of a line for the same voltage and power transfers.

2

1000MW

15

_ q

05

m

>

1 0

0

95

Nid Line Caps

ending End Caps

0 9

0

20

40

60

80 10 0

% Line Length

tom ReceiAng End

Figure 4.3-2 - Line voltage

profile for series capacitors

at the end and middle of a line

The line lengths - and by extension the line impedances - used for the study are, of necessity,

preliminary since the routes of the lines have not been finalized. The changes in final line

impedances will have some impact on the study results since the final routings may increase the

length of the lines. There are several options to address the line length increases:

Maintain a constant net line impedance to ensure the same performance as seen in the

study. As the line length increases, this will require higher levels of compensation and

line voltage profiles will need to be reviewed to ensure that the design criteria are met.

I


32 ERCOT

000054


6/35

Hold the series compen sat ion

to 50% of the

l ine im pedance and run addit ional studies to

determine i f more or larger SVCs are nee ded to provide acceptable system

performance.

Hold the series compen sat ion

to 50% of the

l ine im pedance and run addit ional studies to

determine i f ser ies compen sat ion is needed in other l ines to provide acceptable system

performance.

Shorter line lengths that those used in the study are not a concern since they will have lower

impedances.

4.3.3

Maximum Continuous Current and 30 Minute Overload Ratings

The fundamental ratings for the continuous operating currents and the 30 minute overload

currents were es tablished using the resu l ts of the fundamen ta l f requency study discussed in

Section 3

specif ical ly the generator dispatches and system contingencies that m aximize the

current f lows through the se ries capacitors. With a redispatch

of 10

additional wind gen eration

to maxim ize the l ine loading through the series capacitors, the worst case contingency u nder a

worst case wind dispatch

(

determined from optimal powerflow analyses) established the

maximum series capacitor currents.

Series capacitors are typically designed to have a 30 minute overload rating. This overload

capabili ty is gen eral ly used fol lowing contingencies w here the system can be readjusted within

the 30 minutes to reduce the loading. Since the maximum currents were determined as

discussed above, the 30 m inute ra t ing could be establ ished by these maximu m currents. The

continuous rat ing could be selected to me et norm al system requ iremen ts. This would allow for a

more economical design. However,

the TSPs

may want to have the continuous rating be

established by the m aximu m currents in order to m eet any unknown future requi remen ts.

4.3.4

Maximum Swing Currents

Following a contingency on the system, particularly one that results in line outages, the power

flows through the system will change as the network settles into a new operating condition,

many times experiencing overshoots during the process. The currents associated with these

dynamic swings are temporary, but may be higher than the steady-state maximum currents.

Some dynamic analyses were performed to monitor the highest anticipated swing currents in

the CREZ system. These have been reported to ERCOT and the TSPs for their consideration in

rating the series capacitors.

4.3.5

Maximum Fault Currents

The maximum fault currents through the series capacitors are also an important consideration

for the design of the capacitor protections. The location of the faults relative to the series

compensated line must be considered. Those faults that occur on the line with the series

capacitor segment being considered are known as internal faults, while those not on that line

E3800-PR-00


33

^^5i


7/35

are called external faults. The maximum fault currents determined from a protective case and

also from the various power flow scenarios (e.g. Initial Build, Maximum Edison, etc.) were

determined for both internal and external faults for all series capacitor segments. The results

have been reported to ERCOT and the TSPs for their consideration in the series capacitor

protection design.

4.4

Network Challenges with Series Compensation

Series capacitor compensation has been used successfully in many locations around the world,

and is a relatively common feature in the transmission systems of the utilities in the west and

southwest U.S. However, the resonances that occur between the series capacitor, the

transmission system and electric machines have the potential to result in catastrophic failure of

the machines. Because the series capacitors are always selected to compensate only a portion

of the transmission line of which they are a part, these resonances will always occur at

frequencies below the normal system frequency - in other words, at subsynchronous

frequencies.

Regarding such subsynchronous resonances (SSR) with conventional thermal generators, the

phenomena is well understood and the issues can generally be avoided by judicious design of

the transmission system, by operation of the system around conditions leading to problems

and/or by protection of the machines when undesirable resonant conditions are detected.

With regard to the resonances with wind generation, some events have been experienced and

the industry is quickly gaining a fuller appreciation for and understanding of the phenomena

involved. Papers are becoming more common to address aspects of the issues and to propose

some methods of mitigation, but as of the date of this report, no solution has actually been

implemented and fully tested in the field.

Nevertheless, because the CREZ transmission plan includes multiple series capacitors, ERCOT

and the TSPs considered it prudent to include evaluations of the phenomena to estimate their

potential for occurring on the CREZ system and to test (via simulation) various mitigation

methods. The follow sections describe this work.

4.4. 9

SSI with Wind Generation

While the potentially detrimental, series capacitor

related phenomena evaluated in this study are

associated with

subsynchronous resonances,

they do not always appear to be solely associated

with the electrical resonance itself. In some

Subsynchronous Interactions

Wind

Generators

vs.

Series

apacitors

r mp lr f r ed

Full

CREZ

Test

System

Transmission

syste

cases, they appear to be exacerbated

bSysem. J

controls for the power e lectronic conver ters use d on som e types of wind turbine gen erators.

Because the causes m ay be more gener ic than just the subs ynchronous reso nance, the term

I


34

ERCOT

56


8/35

subsynchronous interaction (SSI) was selected for use in this report to discuss the phenomena

affecting wind generation. Specifically, the following types of SSI are considered

*

Self-excitation - a phenomenon that occurs because of the natural response

(resonances) of the various system components to each other. It is typically stimulated

by some system perturbation; and,

Control interactions - phenomena that occur, in part, because of the response of

active system devices such as the WTG controls.

The phenomena leading to different types of SSI can be complicated given the complexity of the

controls used in some types of wind turbine generators. Because of this, the SSI issues with

WTG were first evaluated with the wind farms connected to a simplified radial test system and

then confirmed on the full interconnected CREZ system.

The simplified radial test system is illustrated in Figure 4.4-1. This topology is most susceptible

to SSI and allowed a more rapid assessment of the issues. Tests were made representing each

of the different types of wind turbines at the wind farm collector bus. They were started with the

series capacitor bypass breaker closed and their susceptibility to SSI was tested by simply

opening the bypass breakers. This was generally enough of a disturbance to trigger any

interaction.

Wind farm

Transmission

collector

line

bu s I r7c I

_ Y

Series

Network

capacitor

equivalent

34.5kV

138kV

Transmission

Strong,

medium

Y

line

or weak

bypass

220 miles

138kV

breaker

or 80 miles

345kV

Figure 4.4 1: Simplified radial test system for SSI evaluations

The confirmation of the test system results on the full interconnected CREZ system, with wind

farms at the locations currently projected by ERCOT, permitted an assessment of the likelihood

for SSI at these locations.

WTG Types

Four basic types of wind turbine generators have been identified in the industry based on their

configuration and operation. These four types are:

Type 1 is a fixed speed wind turbine connected to an induction generator that is, in turn,

directly connected to the grid.

E3800-PR-00

CREZ R eact ive Power Co mpensat ion S tudy

35

57


9/35

Type 2 is a variable speed wind turbine connected to a wound rotor induction generator

which has a controlled variable external rotor resistance that is used to increase the

operating speed range of the generator.

Type 3 is a doubly-fed induction generator (DFIG) which is also called by some authors

a doubly-fed asynchronous generator (DFAG). It uses a variable-speed wind turbine

connected to a wound rotor induction generator. A back-to-back converter is connected

between the generator rotor and the stator in parallel to the machine. Because the full

machine power does not flow through the converter, it can be rated for only a fraction of

the WTG rating. It has a wider operating speed range than Type I and Type 2.

*

Type 4 is a variable-speed wind turbine with a generator (either asynchronous or

synchronous generator) connected to the grid through a back-to-back converter. The

power of the generator flows directly through the converter so it must be rated for full

generator power. The converter acts to decouple the turbine and generator from many

phenomena occurring on the grid.

Self-excitation

with Type 1 and Type

2 Machines

Several models of Type 1 and Type 2 WTGs were provided by ERCOT for evaluation of SSI

issues. Not unexpectedly, a phenomenon known as self-excitation was observed with these

types of machines under certain conditions on the simplified radial test system. Self-excitation is

a well understood phenomenon that is a direct consequence of the resonance between the

series capacitor and the system and machine inductances on the system. Excellent papers (see

references [2] and [3]) were written many years ago that are still pertinent for understanding the

conditions conducive to self-excitation and that provide insight into how it can be mitigated. The

potential for its occurrence with wind turbine generators was noted in reference [4].

Whether or not SSI was observed on the radial test system strongly depended upon the losses

in the system and the parameters of the particular machine. Higher amounts of resistance in the

system between the wind farm and the series capacitor (due to lower voltage transmission

systems, for example) will decrease the likelihood of any undamped resonance conditions

occurring.

At present, ERCOT anticipates that only about 15% of the new wind turbine generators to be

added to the system will be Type 1 or Type 2. But it is generally recommended that the new

plant owners perform a study to assess the potential for self-excitation of their machines if they

wil l

be connecting in the vicinity of any of the series compensated lines, or if a reasonable

number of system line outages would place their plant nearly radially connected through a

series capacitor.

I


36 ERCOT

00005S


10/35

SSI with Type

3 Machines

Out of the various types of wind turbine generators, Type 3 was found to be the most

susceptible to SSI. This appears to be because of interactions with the controls and the

subsynchronous series resonance.

Only two models were made available for assessment in this study and the susceptibility to SSI

was found to differ between the models. The first, more susceptible model had a more detailed

representation of the converter and its controls. This along with parameter differences may

account for its greater susceptibility to SSI.

Because the Type 3 machines' high susceptibility, and because ERCOT currently anticipates

that a significant portion of the new wind turbines installed in the CREZ may be of this type, they

were carefully tested on a model of the full CREZ system. Representations were made of wind

farms at the same locations that they were represented in the fundamental frequency analyses.

These simulations showed particular inclination toward SSI at specific locations as listed in

Table 4.4-1. This table also indicates the system conditions for which the SSI occurred and how

each of the two Type 3 models responded.

Table 4.4-1: Conditions found to be conducive to SSI with Type 3 WTGs on CREZ system

Wind

Size of

turbine

represented System

generator

wind farm

contingency

Model I

Model 2

#

location

M

conditions

Case description

SSI

SSI

I

West

74 3

N-0

Normal system conditions

Y N

Shackelford

2

West

74 3 N-1

Ou tage of one circuit of the double circuit

no t

tested

Y

Shackelford

l ine between Scurry and West Sh ackelford

3 West

74 3

N- 2

Ou tage of double circuit l ine between Scu rry

not tested Y

Shackelford

and West Shackelford

4 West

743 N-2

Ou tage of double circuit l ine between West

Y

not tested

Shackelford

Shackelford and Romney

5

West

743

N-2

Outage of double circuit line between Clear

Y

not tes ted

Shackelford

Crossing and West Shackelford

6

Big Hill

15 0

N-1 ' Outage of circuit between Big Hill and Twin

Y

N

B u t t e s

7 Big Hill

15 0

N-2

Outage of circuits between B ig Hill and Twin

Y

Y

Buttes and between Big Hill and Bakersfield

8

Dermott

56 1

N-2

Outage of double-circuit l ine between

Y N

Dermott and Scurry

9

Dermott

56 1

N-4

Outage of double-circuit l ine between

Y Y

Derm ott and Scurry and doub le-circuit l ine

between Dermott and Cottonwood

Without mitigation measures, there is a strong potential for SSI with Type 3 wind turbine

generators located very close to the West Shackelford, Big Hill and Dermott buses. The first

E3800-PR-00

CREZ R eact ive Power Compensat ion

Study

37

090-059


11/35

Type 3 model, in particular, showed vulnerability at these locations with SSI being observed at

West Shackelford with no line outages.

Because the models assessed in the study are not representative of all WTG manufacturers

and may not provide sufficient detail needed for a full assessment under the studied conditions,

these results should be taken primarily as a caution and detailed studies should be conducted

by the developers to ensure that the planned wind farm will not have SSI issues. Such studies

should accurately represent the CREZ system actually built, any system level mitigation applied

and any WTG level mitigation available from the manufacturers and included in the turbines

being ordered.

Several potent ia l mi t igat ion m ethods, thei r ef fect iveness and their l imi ta tions are discussed

below

in Section 4.4.3.

While the simulations performed for the study can be considered somewhat theoretical, there is

actual experience that emphasizes the importance of the recommended studies. A utility on the

ERCOT system reported an incident in which a wind farm consisting of Type 3 wind turbines

was radially connected to a series compensated line following an N-1 contingency. The

response of the wind turbines to the new system conditions with a more direct influence from

the series capacitor resulted in the tripping of the wind turbines, but not before equipment had

been damaged. It has been reported that the damage was not limited to the WTGs themselves,

but that the series capacitor also sustained some damage. Because of this experience, two

recommendations are made regarding the protection of the series capacitors: 1) interconnection

studies for new wind farms should include an evaluation of the potential for SSI and the

anticipated impact on voltages at and currents through the CREZ series capacitors; and, 2)

design efforts for the CREZ series capacitors should include an evaluation of the impact of

various levels of subsynchronous currents, with protection schemes and/or SSI mitigation added

if

warranted by the evaluation results.

Type Machines

In the evaluations m ade for this study,

the Type 4 WTGs

were n ot found to be affected by the

prese nce of the series capacitors on the system. This

is bel ieved to be due to the decoupl ing

that the full back-to-back converter provides. Although not observed here with the limited

num ber of m odels available for assess me nt, i t is theoretical ly possible that some control issues

could occur. The evaluat ion into any such issue s is left to when they man ifest themselves.

4.4.2 SSR

with Thermal Generation

Subsynchronous Resonance (SSR) is a well-

known phenomenon in which a series resonance

between a generator and a series compensated ac

transmission circuit can destabilize one or more


38

ERCOT

Subsvnchronous Resonance

Thermal enerators and

Series

apacitors

000060


12/35

torsional

modes of oscillation on a generator shaft. Since the discovery in 1971 of the SSR

problem on the Mohave generators of the Southern California Edison Co. there has been only

one event where damage occurred because of the SSR problem with series compensation.

Because of the partial - but high percentage - compensation (63%) of the 500kV line between

the

Mohave plant and the Lugo Substation, when a short line between the McCullough

Substation and Mohave was opened, the system was tuned to a torsional mode involving the

shaft between the generator and a directly-connected exciter. Other instances of SSR have also

occurred at other locations, but damage to the generators involved has been avoided through

proper mitigation or protection methods.

Because the proposed

C R E Z

t ransm ission includes many s eries capacitors,

ERCOT

has taken

a prudent step and asked ABB to perform an SSR scre ening analysis to assess the potential of

SSR between the CREZ transmission and several nearby thermal generating plants. These

screening analyses have considered both the poten tial for SSR and for the induction gene rator

effect (

self-excitation involving on ly the electrical aspects o f the syste m ).

The S S R

Phenomena

In

order to understand the SSR phenomena as it relates to conventional thermal plants,

consideration must be given to both the torsional modes of turbine-generators and the electrical

resonance created by the series compensated line.

Generator Torsional Modes

As discussed in Section 3.7, a mechanical system with N masses with have N-1 oscillatory

mechanical torsional modes. Consider again the generic turbine-generator system as illustrated

in

Figure 3.7-1 and repeated in Figure 4.4-2. In this case there are six masses - the high-

pressure and intermediate pressure turbines, the two segments of the low-pressure turbine, the

generator and the exciter. Any given system may have more or less masses on the shaft.

The frequency of each oscillatory mode and how well it is damped (decays away) will be

dependent upon the relative sizes of the masses, the stiffness of the shaft and the magnitude of

various losses in the mechanical system. Of these modes, those that occur at frequencies

below the system frequency - in other words, at subsynchronous frequencies - are of particular

concern.

Rotating

HP-1P Turbine

LP Turbine

Generator

Exciter

Figure 4.4-2: Generic Turbine-Generator System

The various masses on the shaft will have different degrees of participation in the different

modes. The modes in which the generator itself has significant participation will be more

E3800-PR-00

CRE Z React ive Power Com pensat ion S tudy

39

a0-?D0Gi


13/35

susceptible to SSR. For these modes, a disturbance of the electrical system, such as a fault, will

cause a corresponding torsional disturbance on the generator which is translated to the shaft by

the machine which will also disturb the mechanical system. This will cause the masses to

oscillate against each other at their various natural frequencies, with some modes stimulated

more than others.

The mechanical system always acts to damp out these oscillations over time (i.e. it is positively

damped). The amount of mechanical damping is higher when the generators are fully loaded

than when they are at minimum load.

Subsynchronous Resonance SSR)

For the modes in which the generator participates, currents associated with the mechanical

mode oscillations will be generated and injected into the electrical system. The electrical system

will

usually provide positive damping against these currents, but under proper conditions

negative damping can result. If electrical system damping is negative but is not sufficient to

completely overcome the damping of the mechanical system, then the oscillations will simply

take more time to decay, which is not usually a concern. However, if the electrical system

provides enough negative damping to overcome the positive mechanical damping, then the

oscillations will grow and, if proper protection is not applied, can result in catastrophic damage

to the turbine generator.

The conditions leading to negative electrical damping can be set up,with series compensation

system such as that in Figure 4.4-3. In this figure, the resistance of the elements and the details

of the generator flux dynamics are ignored for simplicity.

Infinite

Bus

X d

X T

X S

XC

Figure 4.4-3: Example series compensated network

This e lectr ica l network consists of the induct ive gene rator sub-transient reactance

(Xd ) , the

inductive transformer leakage reactance

XT

the inductive line reactance

Xs and the

capacit ive series com pensation reactance

(Xc) . Therefore,

the total inductive reactance is

XL=Xd+XT+Xs

A series res onance resu l ts with the combinat ion

of

XL

and Xc

at a frequency of

T

Al - f0

where fo is the normal system frequency (60Hz)


40 ERCOT

ooo sz


14/35

Because se ries compen sation is not designed to fully comp ensate the entire transm ission l ine

(not to mention any transformer or generator reactance)

Xc

will always

be less than

XL

and the

resonan t frequency wil l be below

fo .

If f

is at or near the subsynchronous sideband frequen cy

associated with the curren ts injected into the system due to the m echanical osci llat ions, then

energy can readily transfer between the m echanical and electrical system s.

From

the rotor side

of the machine these frequencies will result in apparent resistances in the machine that are

negative and which can overcome the posit ive resist ive losses of the electrical system.

This will

cause the e lectrical system to provide negative dam ping on the turbin e-gen erator shaft. I f this

negative damping is large enough to overcome the mechanical damping, then the torsional

mode becomes destabilized and oscillations at the modal

frequency

will

be sustained

indefinitely or grow.

Such SSR has historically been a problem primari ly for large steam generators. A gen erator that

is connected radial ly to a highly series-com pens ated transm ission l ine can be at considerable

r isk for undamped subsyn chronous os ci lla tions.

The risk

also exists for generators in an

interconnected network, al though to a lesser degree for highly meshed s ystem s.

Induc t ion

Generator

Effect

The induction generator effect is also associated with the subsynchronous resonances of the

machine with the network. However, it involves only the electrical network and not the

mechanical system. At frequencies below the nominal system frequency, synchronous

generators appear as induction machines, so the same phenomenon that results in self-

excitation of induction generators discussed above can occur. However, this effect is usually

called the Induction Generator Effect (IGE) when speaking about synchronous machines.

Fortunately, the same analysis used to screen for SSR, as discussed next, is ideal for

evaluating induction generator effect.

SSR Screening Analyses

Analyses were conducted for selected thermal plants in the

ERCOT system

near the series

compe nsated l ines to screen

for the likelihood of SSR. The six

plants that were screened are:

Comanche Peak nuclear plant

.

Hays combined cycle plant

Odessa-Ector combined cycle plant

.

Oklaunion coal plant

.

Tradinghouse coal plant

.

Willow Creek combined cycle plant

Screening methods based on frequency scans of the network impedance from behind the

generator under study can be made based on principles discussed in [5]. Care must be taken to

E3800-PR-00

CREZ

Reactive Power Compensation

Study

41

^^^GIG-3


15/35

adequately represent the system components so that their influence is properly taken into

account. In particular, the representation of the loads and generators, including that of multiple

units, is essential. The frequency scan approach for SSR screening is limited to a one-machine-

at-a-time approach. Therefore, when studying multiple units at a common high-side bus care is

required in interpreting and handling the data.

In

addition, the scans must be made under multiple system conditions. Under contingency

conditions, the outages of lines may result in the generators being more directly coupled to the

series capacitors increasing the potential for SSR. Outages can also cause the frequency of the

resonance to shift, aligning it with a generator mechanical mode that was not previously at risk.

A large number of outage conditions were considered for each studied plant to consider all

conditions from normal operation with all lines out, to a direct radial connection between the

studied generator and the nearby series capacitors.

A separate report for each plant has been provided to ERCOT. The reports will be provided by

ERCOT to the individual plant owners. The data and results of these studies contain protected

confidential information and may be considered Critical Energy Infrastructure Information. They

will, therefore, not be made publicly available.

It

was noted during the study that the frequency dependent nature of the impedance presented

by the WTGs to the system is critical to the proper screening for SSR and proper calculation of

induction generator effects at the thermal generators. The representation of Type I machines is

straight forward. Type 2 can become somewhat more complicated but is expected to be similar

to Type 1. Representations for Type 3 and Type 4 must be derived from models of WTG

operation. It is recommended that WTG suppliers be required to provide the impedance

characteristics of their machines when looking into the wind farm from the system. These

characteristics should cover a frequency range of 0Hz to 120Hz in 1 Hz or smaller increments for

normal screening studies. Higher frequencies may be needed for other types of harmonic

impedance calculation studies and should also be provided up to approximately I kHz.

4.4.3 Potential Mitigation Measures and Their Limitations

Because of the severity of potential SSI (including SSR) issues, three potential mitigation

methods were evaluated:

Thyristor Controlled

Series

Capacitors

(TCSC). This is

an active device that uses a

thyristor controlled reactor in parallel to the series capacitor.

The TCSC

controls can

regulate how the capacitor appears to the system. This allows the

TCSC to be

used for

other purposes such as to help damp out large area power swings or m ake a given

capacitor appear to have more capacitance at norm al system freque ncies

( i .e. boost).

With proper controls

(

see below) i t is pos sible for the

TCSC to

appear as an inductor

over mos t of the subsynchronou s frequen cy range, thereby el iminat ing m ost concern for

SSI issues with both wind and thermal gen eration

I ABB Grid Systems Consult ing

42

ERCOT

000064


16/35

Series capacitor bypass

filters.

This is a passive device placed in parallel to the series

capacitor. It allows an alternate path to currents at frequencies other than those at the

normal system frequency 60Hz). This changes how the series capacitor appears to the

system at the subsynchronous frequencies.

Two philosophies can be used for selecting the parameters of these filters. The first

( damping-type ) focuses on damping undesirable currents so it increases the system

resistance at subsynchronous frequencies. This can be tailored to focus on specific

issues or frequencies.

The second philosophy ( preventive-type ) focuses on preventing undesirable currents

by making the series capacitor appear inductive over much of the subsynchronous

frequency range, eliminating most concern for SSI issues with both wind and thermal

generation.

W TG

control

modifications.

This is limited to the Type 3 wind turbines. If any SSI

issues were to be found with Type 4, this would also be an option.

The effectiveness of the first two solutions was evaluated

for

Type 3 WTGs in the full

interconnected

CREZ

system for many of the system condit ions that led to SSI as discussed in

Section 4.4.1. The resu lts are shown

in Table 4.4-2

As can be seen by com paring

Table 4.4-2 to Table 4.4-1,

the preventive

type

bypass fi l ter and

the TCSC were

effective in addressing the SSI issue s for the

Type 3 wind

turbines evaluated in

the stu dy. The last condi t ion

(

N-4 outage at Dermott) represented

a very

weak system and

control issu es becam e a problem during the simu lation so the effectiveness is undeterm ined in

this case.

The TCSC

or a preventive bypass

filter

with

s imi lar s ubsynchronous impedance

characterist ics was also found through the SSR screening studies to be effective in el iminating

concerns for SSR when universal ly appl ied.

The damping type bypass filter was not found to be effective by itself. However, in combination

with control modifications at the WTG it may be more effective. It is also noted that an

exhaustive effort was not made to determine the optimal designs of the bypass filters. It may be

possible that a design not evaluated would show greater effectiveness than that shown here.

E3800-PR-00

CREZ

Reactive Power Compensation

Study

43

000065


17/35

Table 4.4-2: Conditions found to be conducive to SSI with Type 3 WTGs on CREZ system

Wind

Modell

Model 1

Model 2

turbine

System

SSI with SSI with SSI with

Model 1 Model 2

generator

contingency

filter

filter filter

SSI w i th

SSI w i th

# location

conditions

Case description type 1*

pe 2**

pe 2**

TCSC

TCSC

I

West

N- 0

Norm al system conditions

Y

N

not tested N

not tested

Shackelford

2

West N-1

Outage o f one circuit of the

not tested

not tested

not tested not tested

not tested

Shackelford

double circuit l ine between

Scurry and West

Shackelford

3

West

N- 2

Outage o f double circuit l ine

not tested

N

N

N

Shackelford

between Scurry and West

Shackelford

4 West N-2

Outage of double circuit line

Y

no t tes ted

no t t es ted N

no t tes ted

Shackelford

between West Shackelford

and Romney

5 West

N-2

Outage of double circuit line

Y

not tested not tested N

no t t es ted

Shackelford

between Clear Crossing

and W est Shackelford

6

Big Hill N-1

Outage of circuit between

Y

N

not tested

not tested

Big Hill and Twin Buttes

7

Big Hill N-2

Outage of circuits between

Y

N

N

N

N

Big Hill and Twin Buttes

and between Big Hill and

Bakersfield

8

Dermott

N-2

Outage of double-circuit

Y

n o t

tested

no t t es ted

N

no t t es ted

l ine between D ermott and

Scurry

9 Dermott N-4

Outage of double-circuit

Y

Weak

N

Weak N

l ine between Dermott and

system

system

Scurry and double-circuit

control

control

l ine between Dermott and

issue

issue

Cottonwood

* - da m pin g ty pe filt er * * - p re ve nt iv e ty pe

The following sections provide brief discussions on the various technologies

T

As i l lus trated in Figure 4.4-4, the

TCSC

consis ts of series capacitors in parallel with a thyristor

controlled reactor that can boost the voltage across the series capacitors and make the

combination appear as a larger capacitive impedance at fundamental frequency. For example,

the fixed series capacitors may have an impedance of 20% of the line impedance and the

thyristor control led inductor can inject a cu rrent that w il l boost the vo ltage by a factor of three,

allowing the

TCSC to compensate

60% of the

l ine reactance.


44

ERCOT

0 0 0 0 6 6


18/35

i v

i L

i C

+ uC -

Figure 4.4-4: TCSC scheme

During TCSC operation, the line current remains almost purely sinusoidal with little distortion

caused by thyristor switching. Near each of the zero crossings of the capacitor voltage the

thyristors are fired to provide a current pulse that circulates through the TCSC capacitor and

inductor causing an increase in the capacitor voltage during the current pulse. The boost

provided to this voltage (i.e. boost factor) can be adjusted by regulating the timing of the

thyristor switching. This boosted voltage with the given line currents presents an effective

impedance to the system that is larger than the fundamental frequency impedance of the

capacitor itself.

In the design considered here, the TCSC boost factor can typically be adjusted between 1.0 and

3.0. The magnitude of the line current is dependent on the total power flow (real and reactive)

on the transmission line. The magnitude of the thyristor current is dependent on the boost level

setting.

The TCSC modeled in this study uses a specially developed Synchronous Voltage Reversal

(SVR) control to determine the firing of the thyristor valve. The SVR control strategy eliminates

any series resonance in the subsynchronous range between the inductor/valve and the series

capacitors.

With the SVR, the effective impedance presented by the TCSC to the system is

inductive over most of the subsynchronous frequency range, which naturally eliminates SSI by

eliminating the subsynchronous resonances between the system and the series capacitor.

Figure 4.4-5 shows the effective TCSC impedance. See reference [6] for a more complete

description of how SVR results in this effective impedance characteristic.

The effective impedance of the TCSC as modeled in the study has an inductive impedance for

frequencies below 42 Hz, meaning that the TCSC is inductive rather than capacitive over most

of the subsynchronous frequency range, while it is capacitive at fundamental frequency.

At

frequencies lower than this 42 Hz cross-over frequency the TCSC is not capacitive and cannot

create a series resonance. This characteristic can eliminate SSI and even has the ability to

mitigate most concerns for subsynchronous resonance (SSR) with thermal generators. There

are multiple installations of TCSCs operating successfully around the world. To date, however,

none have been explicitly applied to address SSI with wind turbine generators. Note also, that

there is a patent pending on the SVR control. It is not known what methods the various vendors

may have available to provide similar performance.

E3800-PR-00

CREZ R eactive Power Com pensat ion S tudy

45

000067


19/35

virfual

reactance

ideal

SVR

t r a n s i t i o n

frequency

/

nd

power flow

control

/frequency band

tN stator

f requency

_rotor

Increasing

boost

level

Ned

capacitor

Figure 4.4-5: TCSC virtual impedance with the SVR control scheme

At fundamenta l f requency,

the TCSC will provide an impedance equivalent to that of the

conventional series capacitors. For the proposed design w ith a 1.2 boost, the actual im pedance

of the capacitors in the

TCSC will only be 83.3% of

the conven tional ser ies capacitors. The

TCSC

capacitors wil l need to be rated for the maxim um l ine current plus the peak current f rom

the thyristor circuit. Due to this curren t from the thyristor circuit, the TCSC capacitors will also

have a higher voltage for which the capacitors will need to be designed.

Series Capacitor Bypass Filter

The basic topology for the bypass

filter

is

shown in Figure 4.4-6. The

filter consists of a

capacitor/inductor

(

Cf and Lf) paral lel combination w ith a series res istor (

Rd) for damping. RL is

resistan ce of the inductor coil and Cs c is the series capacitor itself.

Csc

Isc

Figure 4.4-6: Series capacitor and bypass filter configuration

There are different design philosophies that can be pursued for a series capacitor bypass filter.

The first is based on the classical solution to self-excitation - providing sufficient damping to

prevent the phenomenon or to cause it to decay before it becomes critical to system

I ABB Grid Systems Consulting

46 ERCOT

000068


20/35


21/35

thousands (drives) of reliable installations around the world by multiple manufacturers

that have been operating for many years - even decades. In fact, most major

manufacturers who have the ability to supply TCSCs state that they have functioning

installations around the world that are operating reliably.

There are only a couple of known instances for which a

TCSC has been applied

specif ical ly to address su bsynchronou s issu es. In one of these instal lat ions, the TCSC

was us ed to address SSR with thermal generators by spl i t ting the ser ies capaci tor so

that part was f ixed and part was

TCSC.

This adjusted the reson ant frequen cy so that it

was not a con cern for any torsional mode on the m achine. Since no instal lat ion exists to

specifically address SSI with wind turbine generators, the beneficial characteristics of the

device have been demonstrated only by engineering calculation and in simulation.

Because of this, the confidence of potential owners of the technology is somewhat

m uted. Further, the potential owners would l ike to have a guarantee that the technology

will eliminate SSI issues, but manufacturers are hesitant to accept the liability associated

with such a guarantee given the novel ty and l imi ted understanding of the phenom ena

involved.

Prices for a

TCSC have

been reported to be around 1.8 times that of a conventional

series capacitor of the sam e rat ings

although one man ufacturer reported a price of 4 to

5 t imes that of a convent ional series capacitor.

Bypass filter

Like the TCSC, the bypass f i lter is covered by patents ( albeit by a differen t

equ ipmen t manu facturer) that may l imit the num ber of suppl iers available to the potent ial

owne rs, who have been hesi tant to accept a technology l im i ted to a s ingle su ppl ier . It

el iminates the opportunity for a compe ti t ive bid process and increases the risk of l imited

future support for the equipm ent.

While the bypass filter has the advantage of using passive elements, there are no known

installations for SSI/SSR mitigation, so any evaluations to date are largely academic

exercises. As indicated above, the evaluations performed for this study have shown the

preventive

bypass filter to provide adequate performance, but the equipment parameter

calculations show that the filter capacitor is as large, or nearly so, as the series capacitor

itself and very high circulating currents are needed, resulting in very large filter reactors

that must have very low losses (i.e. high Q). The magnetic field clearances needed for

the reactors may significantly increase the land area required.

For the

damping

bypass filter,

the componen ts can be m uch sm aller and resul t in lower

losses in the f il ter. However , as shown above, i t may not be able to address SSI issues

with

WTGs

by itself. If used for this purpose it would likely have to be coupled with

another solut ion such

as

WTG

control

m odifications, thereby div iding the solut ion

between a system level solution and a local development level solution. It can be

observed here that this type of split solution may prove challenging in several areas


48 ERCOT

000070


22/35

including the coordination between the different technologies and allocation of the

mitigation responsibility.

For either the preventive bypass filter or the damping bypass filters, some TSPs have

indicated a strong preference for any supplier to guarantee performance in alleviating or

mitigating SSI issues.

The cost for the bypass filters was available from only one manufacturer and only for a

damping type filter. This manufacturer suggested that the price for a series capacitor

with a damping bypass filter would be 1.5 to 2 times that of a conventional series

capacitor.

.

WTG control modifications As of the date of this report, there are no know installed

and field tested Type 3 WTGs with control modifications that have been designed to

address SSI. It is known that significant work is being performed in both industry and

academia to address this issue and the reports appear promising. However, unless any

successful control modifications can address SSI alone, it may prove necessary to

couple the solution with other partial solutions such as a damping bypass filter. Again,

this would divide the solution between a system level solution and a local development

level

solution.

Coordinating the different technologies and determining the proper

allocation of mitigation responsibility may prove to be difficult.

Unless m ult iple man ufacturers are able to address the issue, patent issu es m ay present

a sim ilar problem to that indicated above for the

TCSC

and bypass fi l ters.

It

would not be unexpected for any manufacturer to have an additional charge on each

WTG that has the SSI mitigation controls, but it is not possible to estimate what that

additional cost may be.

At the time of writing, it appears that one manufacturer has successfully implemented

control modifications that allow operation of their Type 3 turbines at the end of a radial,

series compensated transmission line. It is not known how robust the solution will be for

application at other sites. The solution may prove to be dependent upon the specific

system parameters for this interconnection, but the results are quite encouraging.

The concerns expressed in regard to the potential patent issues noted for each of the

technologies could be alleviated if the patent holders demonstrated a willingness to license the

technology in a manner that would allow others to supply it at a competitive price. While rare,

this is not an unknown practice that has had the benefit of opening up a very large market that

benefited multiple vendors instead of limiting it to a much smaller niche market.

E3800-PR-00

CREZ R eact ive Power Co mp ensation S tudy

49

000071


23/35

5

Conclusions

The study documented in this report is the first of its kind on the ERCOT system concerning the

CREZ transmission and has resulted in several key findings that are summarized below.

Reactive compensation requirements

Series compensation of 50% is required on six 345kV double-circuit transmission lines (12

circuits total). The locations of the series capacitor segments along the length of these lines

as studied were provided by ERCOT and the TSPs. The ultimate locations on the lines will

be established by the TSPs based on maintenance needs, line design criteria and similar

considerations. The locations on the lines will not influence the reactive compensation

requirements.

Shunt compensation is required in a number of different forms. The recommended sizes and

locations for switched shunt reactors have been identified. These reactors are required to

keep bus voltage at acceptable levels under conditions with low power flow on the CREZ

system. The reactors are needed at the time of the initial build of the system.

In addition, the recommended sizes and locations for switched shunt capacitors needed for

voltage support during periods with large amounts of wind generation have been identified

for both the initial build of the CREZ system and for the long term build out envisioned in the

study assumptions. The levels required for the initial build are significantly less than those

for the ultimate build out.

Finally, the size and locations for dynamic reactive compensation have been identified for

both the initial CREZ build and the long term plan. Due to the higher levels of wind

generation in the long term plan, the dynamic reactive requirements are significantly higher

than for the initial build. The dynamic reactive devices must be able to provide continuous

voltage control and respond in less than 50ms, which is well within the capability of devices

such as SVCs and STATCOMs.

Specific assumptions were made regarding the reactive capability and performance of the

CREZ wind farms. Simulation results confirm that the success of the proposed

compensation strategy relies on the availability of reactive support from wind generation as

modeled. This, in turn requires operation of the system with such availability in mind.

Specifically, the support from the wind farms must be available when needed, in the

required quantity and with the required speed suggested by the simulation models. Further,

the system must be operated to allow the wind farms to provide as close to zero reactive

output as possible (to preserve their reactive capability for disturbances), while maintaining

overall high voltages. Extensive testing and monitoring of wind farms is recommended to

ensure that such performance is provided.


5 0

ERCOT

000072


24/35

The potential for subsynchronous torsional interactions (SSTI) between the dynamic

reactive compensation devices and nearby thermal generators has been explored for the

thermal generators closest to the recommended locations of the initial CREZ build out. It is

possible for such interactions to lead to severe damage of the generators. The results

indicate that it should be possible to design the controls of the dynamic shunt devices to

eliminate any detrimental SSTI.

Potential concerns for operation near series capacitors

There are several issues of which generation developers should be cognizant when

operating generation near series compensated lines.

SSI with wind turbines: The first relates to wind farm and has been identified in the report

as subsynchronous interactions (SSI). Type 1 and Type 2 wind turbine generators can

experience self-excitation with the series capacitors that may result in the turbines being

damaged or being tripped off line under protective action. Type 3 (DFIG) machines are more

sensitive to SSI, apparently due to the influence of the controls responding to the

subsynchronous series resonance. Type 4 (full converter) machines have not shown any

sensitivity to SSI in this study.

Type 4 (full converter) machines have not shown any sensitivity to SSI in this study.

The locations in the CREZ system to which wind turbine generators are most likely to be

affected by SSI have been identified.

While the simulations performed for the study can be considered somewhat theoretical,

there is actual experience that emphasizes the importance of the recommended studies. A

utility on the ERCOT system reported an incident in which a wind farm consisting of Type 3

wind turbines was radially connected to a series compensated line following an N-1

contingency. The response of the wind turbines to the new system conditions with a more

direct influence from the series capacitor resulted in the tripping of the wind turbines, but not

before equipment had been damaged. It has been reported that the damage was not limited

to the WTGs themselves, but that the series capacitor also sustained some damage.

Because of this experience, two recommendations are made regarding the protection of the

series capacitors: 1) interconnection studies for new wind farms should include an

evaluation of the potential for SSI and the anticipated impact on voltages at and currents

through the CREZ series capacitors; and, 2) design efforts for the CREZ series capacitors

should include an evaluation of the impact of various levels of subsynchronous currents,

with protection schemes added if warranted by the evaluation results.

SSR with thermal generators: Subsynchronous resonance (SSR) between thermal

generators and series compensated lines has been known since the 1970s. The

phenomena can result in high stresses on the turbine-generator shaft which can lead to

catastrophic results if the turbine-generator is not properly protected. With the introduction of

series compensated lines on the CREZ system, some existing thermal generators may be

E3800-PR-00


51

000073


25/35

susceptible to SSR. Screening studies have been performed on several generators that are

near the CREZ series compensation. These studies were documented in separate reports

that will not be made public because they contain proprietary confidential information and

critical infrastructure information.

A related issue is the so-called induction generator effect that can also result in high levels

of subsynchronous currents in the generators and the connected system. These do not

involve the mechanical system of the turbine-generator shaft.

It

is important for any future thermal generation developers to be aware of the issues

surrounding SSR so that they can investigate the potential for undesirable resonances as

part of their interconnection studies.

Mitigation methods: A few mitigation methods for SSI and SSR are explored in the study.

Bypass filters across the series capacitor, designed to provide an alternate path to

subsynchronous currents were explored. Two philosophies - a damping filter and a

preventive filter - were considered. The damping filter did not prove alone to be successful

to fully eliminate SSI with wind turbine generators, but may be more successful in

combination with other methods. The preventive filter parameters can be selected to

eliminate SSI and SSR, but could result in a very costly design. There are no known

installations of these types of high power bypass filters for SSI/SSR mitigation anywhere in

the world. Estimates from a single vendor indicated a cost of 1.5 - 2.0 times that of a fixed

series capacitor. The performance of the filters considered was unclear. Patents on bypass

filters may limit the number of

suppliers.

A thyristor controlled series capacitor (TCSC) - especially

one w ith a

so-called SVR

control

- was found

to be very effect ive in el iminat ing SSI

and SSR. TCSCs have been su ccessful ly

deployed in m any areas around the world by several vendors, but only one is known to have

been deployed

specifically

to address

SSR. A TCSC

will be m ore expensive than a s imple

series capacitor. Estimates from various vendors ranged from 1.5 to 5.0 t im es that of a f ixed

ser ies capaci tor . Patents on

TCSC

controls, such as the SVR, may limit the number of

suppl iers that can provide the necess ary performance.

The modification of WTG controls -

particularly

for Type

3 turbines

is another m it igat ion

me thod that is showing promise. I t is known that significant work is being performe d in both

industry and academia to address this issue an d the repor ts appear promising. Howe ver ,

unless an y successful control modif ications can address SSI alone, i t may prove neces sary

to couple the solution with other partial solutions such as a damping bypass

filter. This

would divide the solution between a system level solution and a local development level

solut ion. I t can be observed he re that this type of spl i t solut ion m ay prove chal lenging in

several areas including the coordination between the different technologies and allocation of

the m it igation responsibi l ity. Also, unless m ult iple m anufacturers are able to address the SSI

problems, patent issues m ay limit the numbe r of suppl iers.


52

ERCOT

000074


26/35

Limitation of wind turbine types - at critical locations, the limiting the types of WTGs to those

not susceptible to SSI may be an option. The results of this study (with a limited number of

models) indicate that Type 4 turbines may be able to operate without control modifications at

locations where other technologies may have SSI issues.

Operate around the issue - under some conditions, such as when SSI is only expected

when certain lines near the wind turbines are out of service, it may be possible to utilize

special protection schemes to prevent SSI issues. Such schemes require careful study and

may include tripping wind generators or bypassing the series capacitors. It is noted,

however, that bypassing the series capacitors under contingency conditions is not usually

prudent because the series capacitors generally become particularly important under such

contingency conditions. Further, tripping of the wind farms may not be an acceptable, first

level response to SSI.

Modeling needs

for future studies

This study

has highlighted some of the limitations of the present models being used for

evaluating wind generation. Several of the issue s are highl ighted below based on the types

of studies

for which

they are used.

Fundame ntal frequen cy models: The main issue observed in th is s tudy was the sens i t iv ity

of the models to low short-circuit rat ios betwee n the system strength and the instal led wind

generation. Under these condit ions high frequen cy oscil lat ions

(

somet imes in excess of 10

Hz) were obs erved. I t was not clear i f these oscil lations are a res ult of model ing issue s or i f

they

would actually exist in the s ystem .

Additional work

would be ne eded to confirm which is

the case. If it is found that the phenomenon is a modeling issue, then it is strongly

recomm ended that work be done to improve the models to prevent un warranted conclusions

f rom being drawn based on study resul ts using the m odel .

(

Note that in this stu dy, i t was

determined to address the issue by using

"

place holder synchronous condensers to

increase the s hort-circuit rat ios. I f such an increase is actual ly nee ded, other technologies

may also be available to mit igate weak system s)

Another model ing issue observed in the s tudy was the poor performance of som e dynamic

models

provided by

wind developers to ERCOT. These

m odels were m ost l ikely created by

the wind turbine m anufacturers. It is em phasized that most of the models worked well for the

purposes of the study, but theopoor performance of a few created num erous diff icul t ies.

In the future, develope rs wil l sti ll be requ ired to provide appropriate mode ls for their wind

farms. I t is recom me nded that a set of tests be developed

which

al l future m odels mu st pass

before they are accepted

by ERCOT

Electromagnetic transient models: The evaluation of the potential

for SSI

with wi nd

turbines and series capacitors is currently limited to simulations in electromagnetic transient

programs such as

PSCAD.

The number of available models which wind turbine

E3800-PR-00


53

000075


27/35

manufacturers are prepared to release is very limited. This is a situation that is simply

unsustainable because it is likely that future studies will need to combine appropriate

models of equipment from multiple vendors. It is recommended that the wind turbine

manufacturers develop black-box models that allow the user access to appropriate control

parameters while hiding those controls and parameters that are proprietary. Such models

should

be backed by the vendors as being suitable for evaluations involving

subsynchronous, synchronous and higher frequency studies, with a clear explanation of

their limitations.

Frequency scan models: The SSR screening studies showed that the representation of

the Type 3 and Type 4 impedance characteristics are important for accurate assessment of

SSR and induction generator effects. It is recommended that WTG suppliers be required to

provide the impedance characteristics of their machines when looking into the wind farm

from the system. These characteristics should cover a frequency range of 0Hz to 120Hz in

Hz or smaller increments for normal screening studies. Higher frequencies may be needed

for other types of harmonic impedance calculation studies and should also be provided up to

approximately I kHz.

A number of assumptions have been made regarding the locations and chronological

development of the wind generation. Further items such as real estate availability in substations

(e.g. to maintain required clearances), increased annual maintenance and possible forced

outages are not part of the study. Also, actual experience will likely differ somewhat from the

assumptions made in the study. Therefore, the results of the study should be used as input for

the initial design efforts and as a guide for future planning. If actual experience is found to be

significantly different from the assumptions made in the study, some of the results may need to

be re-examined. If the transmission providers significantly change the location of some reactive

compensation, the impact of the relocation on system performance and stability should be

studied.


54

ERCOT

00007G


28/35

6

References

[1 ]

Pouyan Pourbeik, Anders Bostrom, Bhaskar Ray, Modeling and Application Studies for a

Modern Static VAr System Installation, IEEE Transactions on Power Delivery, vol. 21, no. 1, Jan.

2006.

[2 ] J .

W. Butler, C. Concordia, Analysis of Series Capacitor Application Problems,

AIEE

Transactions,

vol. 56, pp. 975-988, 193 7.

[3 ]

C. F. Wagner, Self-excitation of Induction Motors With Series Capacitors,

AIEE

Transactions,

vol. 60, pp. 1241-1247,1941.

[4 ]

P. Po urbeik, R. J. Koessler, D. L. Dickmander, and W . Won g, "Integration of Large W ind Farms

into Uti li ty Grids (Part 2 - Perform ance Issues)," in

Proc.

2003

IEEE PES General Meet ing,

vol. 3,

July 2003.

[5]

P. M . Anderson, B. L. Agrawal, J. E. Van Ness,

Subsynchronous

Resonance

in Power Systems.

New York: IEEE Press, 1990.

[6 ]

Lennart Angquist, Gunnar Ingestrom, Hans-Ake J onsson,

Dynamica l Per fo rmance

of TCSC

Systems,

CIGRE 1996 14-302

[7]

IEEE S SR W orking Group, "Term s, Definitions and Symbo ls for Subsynchro nous Oscil lations,"

IEEE Transact ions on PowerApparatus

and Systems, vol. PAS-104, June 1985, pp.1326-1334.

[8]

P. M . Anderson, R. G. Farmer, Series

Com pens at ion o f Powe r Sys tem s .

Enc initas, California:

PBLSH Inc., 199 0.

[9 ]

Chong Han, Don E. Martin, Modesto Lezama, Transient Over-Voltage (TOV) and Its

Suppression for a Large Wind Farm Utility Interconnection, in

Proc.

of 1 International

C o n f e r e n c e o n S us t a in a b l e P o w e r G e n e r a t io n a n d S up p ly ( S U P E R G EN 0 9 ) ,

No. S03P039 7,

NanJ ing-China, Apr., 2009 .

[10]

D.

Dickmander, B. Thorvaldsson, G. Stromberg, D. Osborn, Control System Design and

Perform ance Verification for the Chester, Maine Static VAR C omp ensator,"

IEEE Transactions on

Powe r Del ivery,

vol. 7, No. 3, July 199 2.

E3800-PR-00

CREZ R eact ive Power Co mpensat ion S tudy

55

000077


29/35

App endix A - Dynam ic Shun t Co m pensat ion Techn ologies

A.1- SVC Technology

A Static Var Compensator (SVC) is a regulated source of leading or lagging reactive power. By

varying its reactive power output in response to the demand of an automatic voltage regulator,

an SVC can maintain virtually constant voltage for dynamic events at the point in the network to

which it is connected. During steady-state it can also reset itself to minimum output. An SVC is

comprised of standard inductive and capacitive branches that are controlled by thyristor valves

and connected in shunt to the transmission network via a step-up transformer. Thyristor control

gives the SVC the characteristic of a variable shunt susceptance. Unlike mechanically switched

compensation, an SVC can operate repeatedly and is not encumbered by the delays associated

with

mechanical switching. This lets the SVC respond very rapidly to changing network

conditions such as line or generator outage contingencies.

An SVC can have an inductive and a capacitive capability. The algebraic difference between

these two capabilities is called the dynamic range. There are three main building blocks

available to make-up the required SVC capability. These are the thyristor-switched capacitor

(TSC), the thyristor-controlled reactor (TCR) and the harmonic filter (HF). The TSC is a

synchronized on-off device. The TCR reactive power absorption is continuously variable from

zero to its rated value due to phase control of its conduction interval which controls the

fundamental frequency component of reactor current. If a TCR is used, harmonic filters are

usually required to limit voltage distortion in the network to acceptable values. The HF is

capacitive at the fundamental frequency and contributes to the net capacitive output of the SVC.

The HF is normally not switched but fixed to the SVC bus and is therefore often referred to as a

fixed capacitor (FC).

There are three basic SVC configurations as shown in Figure A.1-1. The first consists of a

thyristor-switched reactor (TSR) and a thyristor-switched capacitor (TSC). Since no reactor

phase control is used no filters are needed. The second consists of a TCR and TSC (TCR/TSC)

which may also include a fixed filter capacitor. The third consists of a TCR and a fixed filter

capacitor (TCR/FC). In transmission applications the SVC is coupled to the network through a

step-up transformer and the first two configurations are the most common for transmission

requirements. The required reactive power is measured on the high side of the transformer.

Some manufacturers design SVCs with significant redundancy built in. The control system is

completely redundant and one control system can be taken out of service for maintenance

without interrupting the operation of the SVC. The thyristor valves have extra thyristors in series

to provide redundancy. There is an extensive monitoring system as part of the SVC. For

example, if a thyristor fails its location is noted and logged so that it can be replaced during the

next schedule maintenance. The cooling system is a closed system with redundancy built in

I


56

ERCOT

000078


30/35

including redundant pumps. Due to this redundancy maintenance can be performed on the

cooling syste m w ithout interrupting the operation

of the SVC.

r s F e

;sc

ra a

a

b

T

T

TcR

rJtBr3

c

Figure A.1-1: Common SVC configurations

Variations of the above basic configurations can be made to optimize the SVC system design.

This is especially true for the higher rated SVC applications where it is common to have multiple

TSC branches with overall continuous control achieved with a TCR. Factors influencing the

design include continuous and dynamic ratings, loss evaluation, redundancy requirements,

harmonic generation, audible noise, environmental conditions and area constraints. The

transformer and some SVC branches need not be rated continuously. If the maximum output of

the SVC is only required dynamically during system swings for instance, some SVC branches

and the transformer can be rated on a short-term basis. If reliability and redundancy are

extremely critical, the SVC can be split in two halves with each connected to a separate

transformer secondary winding, one wye and the other delta. To extend the overall capacitive

range or restore the SVC output to within its dynamic regulating range or continuous rating

following systems contingencies, a mechanically switched capacitor bank can be installed on

the high voltage bus. Such a configuration is called a static var system (SVS).

The range of SVC operation is shown by its static voltage - current (V-I) characteristic illustrated

in the following figure.

E3800-PR-00

CRE Z React ive Power Com pensat ion S tudy

57

000079


31/35

IV

Current

V limit

War

restriction 1.3

to 150

War

pu

10 '

2% 1 1 o lll

j o

Inductive

Capacitive

10 r6

Slopes

2 r 6

design point

design point

75

War

-150 War

a t 1 0 2 p u

at 0 95 pu

- 400 kV^ V

Base -

i B a s e

= 100 MVA

operating

voltage

Continuous Operation

EM Restricted Operation

2 . 0

1 . 5

1 . 0

0 5 0 . 0

0 5

0 I H V I p u I

Cap Range

Ind Range

Figure A.1-2: SVC static V-I characteristic viewed from the transformer HV side

The norm al continuou s operating area for an example

SVC is

defined by the shaded area in the

above figure.

This

normal operating area is bounded by the inductive and capacitive

susceptance limits and the minimum and maximum slopes of the voltage regulator

character istic. Operation is al lowed o n a restr icted basis on the capacitive side above a M VAr

l im it and on the inductive side above the maximum continuous voltage where the

TCR current is

l imited with increasing system voltage.

An SVC

can respond

dynamically in 20 to 60

milliseconds.

An example

of large SVC systems

is described in Preve nting voltage collapse

by large SVCs at

power system faul ts

by Ahmed H.

AI-Mu barak, Saleh M. Bamsak, Bjorn Thorvaldsson, Mikael

Halonen and Rolf GrUnbaum

( IEEE paper

# 978-1-4244-3811-2/09). This paper

includes a

single- l ine diagram for one of three +600/-60MVAr

system s and also d iscusses som e control

issues en countered on the relat ively weak system to which they

were connected.

A.2 - STATCaM Technology

The Static Synchronous Compensator (STATCOM) is comprised of a voltage source converter

(VSC) connected in shunt as illustrated in Figure A.2-1. The shunt-connected VSC is based on

converter technology with valves comprised of solid-state switching components with turn-off

capability and anti-parallel diodes. Performance of the STATCOM is analogous to that of a

synchronous machine generating a balanced three-phase set of sinusoidal voltages at the


58 ERCOT

000080


32/35

fundamental frequency with controllable amplitude and phase angle. The device, however, has

no inertia and does not contribute to the short circuit capacity.

vt

T

V5C

VDG

IQ

V:> V.

rsac?Ne power

voc

Ahs^xqs

V,


33/35

special transformer connections, the pulse number can be reduced and the resulting staircase

waveforms more closely approximate a fundamental frequency sine wave, thereby reducing the

need

for filtering.

The following figure illustrates waveform generation from a three-level VSC employing pulse

width modulation (PWM) control. The VSC is coupled to the AC bus through series air-core

reactors. Low pass filters tuned to the switching frequency may need to be connected on the

line side of these reactors. Together they form a low-pass filter such that only the fundamenta

series capacitor compensation requirements

Documents